RayLin

I made a Chrome extension that uses OpenAI tokenizer to track token usage on
Anyone interested in testing it?

Fable 5 IQ 🥇 (offline Test)

It is time to say goodbye

Fable 5 + Ultracode is really insane
It independently triggered the workflow 3 times🤯
I used GPT-5.5 Pro for planning and generated the prompt based on the Fable 5 prompting guide:

Opus 4.8 only 200k context window 😭
Same as Claude 2

> Claude
> Really low usage
> Unstable connection
🤡

Alpha models? #gpt4o is back?
@btibor91

Happy New Year!!! 🎇
2025 Model Release Roundup!!! (first launch dates; Meta excluded)
I’m hoping the number of releases in 2026 can break 100!

Anyone interested in this? I forked Claudia and integrated both the Codex CLI and Gemini CLI!
It can track CLI token usage, usage patterns, and total cost.
You can also directly add Claude code sub-agents and Skills.

I’m not sure if I’m the only one who feels this way, but in most cases Opus 4.5 is actually better than GPT-5.1-Codex-Max-Xhigh.
And Claude has also significantly increased the usage limits for Opus.